|
Title:
|
DEVELOPMENT OF A FLOOD FORECASTING SYSTEM IN HYDROGRAPHIC BASINS BY MEANS OF ARTIFICIAL NEURAL NETWORKS |
|
Author(s):
|
Paulo Bousfield, Cladir Zanottelli, Cátia Ganske, Sidney Schossland, Flávio Lapolli, Lorival Chapuis |
|
ISBN:
|
978-972-8939-23-6 |
|
Editors:
|
António Palma dos Reis and Ajith P. Abraham |
|
Year:
|
2010 |
|
Edition:
|
Single |
|
Keywords:
|
Artificial Neural Networks, Hydrographical Basins, Flood Forecast. |
|
Type:
|
Short Paper |
|
First Page:
|
114 |
|
Last Page:
|
118 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
Aiming to eliminate or minimize some problems caused by floods in hydrographic basins, this paper intends to develop a flood forecasting system. As a technique to solve the flood forecasting problem, it uses artificial intelligence, the learning paradigm of connexionist networks. In order to create the Artificial Neural Network ANN, a database was created through the use of historical series from 2008 and 2009 of the daily pluviometric data (amount of rain accumulated in one day, in millimeters) collected from 21 meteorological stations located in a given Hydrographic Basin. Daily fluviometric data has also been used (river level, in centimeters) collected at the specific location the forecast is wanted. The data has been obtained and made available by the National Agency of Water (ANA Agência Nacional de Águas). In order to obtain an ANN that had a good generalization possibility and aiming at finding the number of neurons in the hidden layers, a number of experiments has been designed using feedforward network, supervised learning and backpropagation algorithm. The success of the experiments was based on the result of the average square error and linear regression of the network forecast obtained results. It was clear that the use of alternative techniques, such as the connexionist learning, in this case the ANNs to be used in flood forecasting systems, is appropriate and is highly relevant to improve the hydrological modeling techniques. The Itajaí-Açu River Basin has been chosen as the area of interest, with flood forecast for the city of Blumenau, in the State of Santa Catarina Brazil. |
|
|
|
|
|
|
